RESUMEN
Early detection of individuals with a high risk of dementia is crucial for prompt intervention and clinical care. This study aims to identify high-risk groups for developing dementia by predicting the outcome of the Mini-Mental State Examination (MMSE), using historical data collected from community-based primary care services. To mitigate the effect of inter-individual variability and enhance the accuracy of the prediction, we implemented a multi-stage method powered by supervised and unsupervised machine learning methods. Firstly, we preprocessed the original data by imputing missing values and using a wrapper-based feature selection algorithm to pick significant features, resulting in ten variables out of 567 being selected for further modeling. Secondly, we optimized hierarchical clustering to partition the unlabeled data into groups by their similarities, and then applied supervised machine learning models to build subgroup-specific prediction models for the identified groups. The results demonstrate that the proposed subgroup-specific prediction models generated from the multi-stage method achieved satisfactory performance in predicting the outcome classes of dementia risk. This study highlights the potential of incorporating unsupervised and supervised learning models to predict high-risk cases of dementia early and facilitate better clinical decision-making.
Asunto(s)
Demencia , Aprendizaje Automático Supervisado , Humanos , Anciano , Algoritmos , Demencia/diagnóstico , Atención Primaria de SaludRESUMEN
Objective: To validate the Chinese version of the Quality of Life (QoL) Patient/Cancer Survivor Version (QOLCSV-C) for measuring QoL in Chinese cancer survivors. Methods: The study followed a seven-step research practice guideline for cross-cultural research instrument validation study including translation, adaptation, and psychometric assessment. A forward- and backward-translation procedure was approached, followed by cultural adaptation and acceptability assessment. For its psychometric properties, its concurrent validity with the Functional Assessment of Cancer Therapy-General (FACT-G) was examined with correlation analysis. The internal consistency (Cronbach's alpha) and item-total and item-subtotal correlations of the QOLCSV-C were obtained. Factor analyses were conducted. Floor and ceiling effects and the discriminant performance of the selected variables on QOLCSV-C score were also examined. Results: The QOLCSV-C was translated from the 41-item QOLCSV with four domains: psychological, physical, spiritual and social well-being. The content validity was excellent (CVI â= â1.00). Time spent to complete the QOLCSV-C was about 10 âmin. The QOLCSV-C was found easy to use, appropriate in length, and reflective of their QoL. The strong correlation between QOLCSV-C and FACT-G indicates a satisfactory concurrent validity (Spearman's rho â= â0.765, P â< â0.001, n â= â205). The overall internal consistency of the QOLCSV-C (Cronbach's alpha â= â0.888) and the split-half reliability (Spearman-Brown r â= â0.918) were excellent. Most of the items show moderate to strong item-total correlation. The exploratory factor analysis revealed a four-factor solution, and confirmatory factor analysis has a satisfactory model fit with indicative items. None of the total scores of QOLCSV-C reveal the floor or ceiling effect. For discriminant performance, variables demonstrating significant between-group differences include sleep quality, pain, fatigue, nausea, physical health, and financial burden. Conclusions: The QOLCSV-C is a reliable and valid instrument for measuring the QoL in Chinese cancer survivors. Future studies can explore the factor structure, gender universal or specific items, and significant predictors of QoL of cancer survivors in different cultures.